Choosing Among Living-Donor and Cadaveric Livers

Authors:
Oguzhan Alagoz;Lisa M. Maillart;Andrew J. Schaefer;Mark S. Roberts
Affiliations:
Department of Industrial and Systems Engineering, University of Wisconsin--Madison, Madison, Wisconsin 53706;Department of Industrial Engineering, University of Pittsburgh, Pittsburgh, Pennsylvania 15261;Department of Industrial Engineering, University of Pittsburgh, Pittsburgh, Pennsylvania 15261;Section of Decision Sciences and Clinical Systems Modeling, Division of General Internal Medicine, School of Medicine, University of Pittsburgh, Pittsburgh, Pennsylvania 15213
Venue:
Management Science
Year:
2007

Citing 0
Cited 4

Eliciting Patients' Revealed Preferences: An Inverse Markov Decision Process Approach

Decision Analysis
A Broader View of Designing the Liver Allocation System

Operations Research
A biologically based discrete-event simulation model of liver transplantation in the United States for pediatric and adult patients

Proceedings of the Winter Simulation Conference
A comparison of decision-maker perspectives for optimal cholesterol treatment

IBM Journal of Research and Development

Quantified Score

Hi-index	0.01

Visualization

Abstract

The only therapy for a patient with end-stage liver disease (ESLD) is liver transplantation, which is performed by using either a cadaveric liver from a deceased donor or a portion of a living-donor's liver. This study addresses the following decision problem for an ESLD patient with an available living donor. Should she have a transplantation now or wait? If she decides to have the transplantation now, should she use her living-donor liver or a cadaveric liver for transplantation? We formulate this problem as a discrete-time, infinite-horizon Markov decision process model and solve it using clinical data. Because living donors are typically related to the recipient, we incorporate a disutility associated with using the living-donor liver as opposed to using a cadaveric liver. We perform a structural analysis of the model, including a set of intuitive conditions that ensure the existence of structured policies such as an at-most-three-region (AM3R) optimal policy. Our computational experiments confirm that the optimal policy is typically of AM3R type.